[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Certificate Cancellation Notices (CCN)



	 At 11:56 AM 4/3/97 EST, Ron Rivest wrote:
	 >The CCN allows for some "softening" of SPKI's "once-issued, no
	 >revocation" policy by enabling an issuer to attempt to control the
	 >extent to which an adversary can make use of a certificate that the
	 >issuer would like to revoke, without having enforceable revocation.
	 >
	 >However, if the issuer can distribute the CCN to the right places (wh
	ich
	 >might only be a few verifiers), he can get close to what CRL's
	 >attempt to achieve, without all of their complexity and costs of CRL'
	s.
	 >CCN's are an informal version of CRL's, and get most of the benefits 
	with
	 >little of the cost, which is what you want in a system that calls its
	elf
	 >"simple".
	 
	 "softening" reminds me of "a little bit pregnant".
	 
	 I believe that if we consider the server to be an online service which
	 is backing up online tests you have specified in your certs and if
	 you find a way to replicate databases between issuer and that server
	 for the purpose of answering such online inquiries, then we have the
	 effect you're looking for without going soft on the definition.

The point of CRLs is to avoid the need for online services.  It's not so
much the replication of the database that concerns me; rather, it's the
requirement that all possible acceptors of certificates be online to do
any processing whatsoever.

Let me give a very real example.  Suppose I download a batch of email to
my laptop, hang up the phone, and then start wading through the messages.
One message contains some signed, active content.  Has the certificate
attesting to it been revoked?  Must I dial up again to finish reading that
message?

Granted, I may not want to have complete CRLs for the email population of
the Internet continually downloaded to my laptop.  But in smaller communities,
that's not an unreasonable requirement.  STU-IIIs are a good example.

Follow-Ups: